All right. Welcome, everybody. On behalf of my co-chair, Helmut Pirfelner, welcome you to this session entitled Unresolved Issues in the Management of Atrial Fibrillation. So welcome to the session. Welcome to Heart Rhythm Society 2025 here in San Diego. And for all those who have not done so, you can download the HRS mobile app from your app store, and that can help you with the Q&A, which we can take either at the microphones, the microphone rather, in the room, or we can take it via the app over the iPad here to answer your questions. So we want this to be interactive. These are all super controversial topics, and we have great speakers, experts in the field that are going to present. Just, you know, we want to remind everyone it's strictly prohibited to make video or still photography of the speakers. If you really want a picture of Bill McIntyre, you can take it in the hallway after the session. He does those for free. So without any further ado, I'd like to invite our first speaker, a great expert and friend, Lena Rivard from the Montreal Heart Institute, to talk about the prevention of cognitive dysfunction in patients with atrial fibrillation. So Lena, welcome to the stage, and everyone else, welcome to the session. Hi. Good morning, everyone. It's my pleasure to be here today. I'm Lena Rivard from the Montreal Heart Institute, and I will try to present on the prevention of cognitive decline in patients with atrial fibrillation with a focus in patients with SCAF. So I'm at disclosure. So we all know here that atrial fibrillation is linked to cognitive decline and dementia, even in the absence of clinical stroke and after accounting for shared co-factors. We also know that the relative risk is higher in younger individuals, meaning less than 65, 70 years old, and that patients with atrial fibrillation have an increased rate of brain lesions, not only ischemic brain infarcts, but also higher rate of white matter hyperdensities and lower brain volume. There are several suspect mechanisms, obviously all the shared risk factors, but also silent and clinical brain infarcts, hyperperfusion, microbleeds, and inflammation. So what are the preventive treatments? Oral anticoagulation, AF ablation, multidisciplinary approach, those are the most commonly thought about. Oral anticoagulation, so we have multiple large observational studies that have reported a decrease of the risk of protective effect of oral anticoagulation in patients with atrial fibrillation, especially when started early in the course of atrial fibrillation. In patients on warfarin, those who are under or over oral anticoagulated are at high risk of cognitive decline. But we have to be careful because obviously all those studies are non-randomized, and the protective effect of oral anticoagulation may be explained by the fact that patients under oral anticoagulation, the population is slightly different compared to those in whom oral anticoagulation is not given. This is a great study published in European Heart Journal that you may know in 2017 from the Swedish registry. So they looked at more than 160,000 AF patients. They look at the risk of dementia after the first diagnosis of atrial fibrillation, and as you can see, the risk is much higher in patients without oral anticoagulation and much lower in patients in whom oral anticoagulation is started early after the first diagnosis of atrial fibrillation. This is another study from the Nippon AF registry in all elderly patients, around 3,000 patients, mean age 81. So at baseline, 20% of the AF population had cognitive impairment defined as MMSE lower than 24. And after two years, one third of the population had a decrease of at least two points of MMSE. They found that even after adjustment, oral anticoagulation therapy was associated with less cognitive decline at two years. And that also, and the second finding of the studies was that patients with cognitive impairment at baseline had higher risk of death, stroke, heart failure during follow-up. For patients with silent atrial fibrillation, we looked at the cognitive function in the sub-study of the Artesia trial, which we presented at EAC last year. So we, in the main Artesia participant, from the main Artesia participant, around 1,000 entered the cognitive sub-study. Cognitive testing was performed at baseline with the MOCA, and we also collected additional parameters linked to cognition. After two years, a repeat MOCA score was performed in 658 patients. So primary endpoint was a change in MOCA score, and secondary endpoint was a two-point reduction in MOCA score compared to baseline, and a change in the visual spatial executive functioning component of the MOCA. What we found in patients with device-detected atrial fibrillation, the MOCA score decreased from 23.8 to 23.7 in the APIC-Sebron group, and from 23.9 to 23.5 in the Aspirin group, which was no different. We found the same result in the modified ITT and the on-treatment analysis. In the on-treatment analysis, what we found was that oral anticoagulation was associated with significantly less decline in visual spatial and executive functioning. In the NOAA study, they also performed cognitive testing in their population. So all the MOCA score was performed at baseline one year and two years of follow-up. After two years, around one-third of the population had a repeat MOCA. They didn't find any difference between placebo and oxaban group, and the visual spatial and executive functioning was not published in this appendix. So oral anticoagulation in SCARF in patients with device-detected atrial fibrillation, oral anticoagulation did not affect the evolution of the MOCA score, and more studies are needed to assess the role of oral anticoagulation in prevention of visual spatial and executive functioning decline. There are also other mechanisms that can be involved, and one of the main suspects is the hippoperfusion of the brain. This is a nice study published in 2018 from Iceland. What they do is they did contrast brain MRI in AF patients. Mean age was 79. Patient did not have any history of heart failure or cognitive impairment. And they found that brain perfusion was lower in patient with atrial fibrillation at the time of the MRI when compared to patients who had atrial fibrillation but were in sinus return at the time, or patients who never had any atrial fibrillation history. So right now, there is several studies ongoing on AF ablation. Most of the studies are longitudinal, obviously. And we have one small randomized controlled trial on 100 patients. Patient cognitive function was compared in patients who underwent AF ablation or not. And what they saw is a transient early post-operative cognitive dysfunction that has been well described in AF ablation during the first three months. But an improvement in overall cognitive function at one year. The integrated approach can also be used in patients with SCARF. It has been shown to be efficient in patients with clinical atrial fibrillation. In the current national wine registry, anticoagulation and good blood pressure control were linked to a decrease in dementia. And in a very recent study from the Swiss AF study, they found that in patients with atrial fibrillation, the number of the brain lesions and the brain volume was linked to smoking duration. I mean, patients with a longer smoking duration have more brain lesions and a smaller brain volume compared to patients who didn't smoke. So what is the future of prevention of cognitive impairment in patients with atrial fibrillation? We must work now on the role of brain hyperperfusion in atrial fibrillation. And we don't know if it's a cause in SCARF. We need to investigate the role of anticoagulation in prevention in decline in visuospatial and executive functioning in those patients and obviously assess cognitive function in all future AF studies. Thank you for your attention. So we have time for a few questions if anyone would like. I'll start off while you're getting to the microphone. Lena, just frame it for us. A MOCA score of 23 out of 30, how compromised is that? Can I go to work if my MOCA score is 23? Can you be an EP? I won't answer this question. But my cognitive impairment is defined by a MOCA score lower than 26. But obviously, the lower you have, the more specific you are. Because a MOCA score, like the NMSE, is just a screening test. You cannot use that. We're using that in studies because obviously it's a 10-minute test. So we cannot perform one hour test in 1,000 patients every year. It's just not possible. But that would be ideal because we cannot define my cognitive impairment just on NMSE or MOCA test. I understand that. But 26 is a cutoff. Premesh Kovu from Sydney, Australia. The current ESC guidelines for blood pressure control aims for quite low blood pressure. I think it's 110 over 70 or something. So my question is, if we go too low with the blood pressure, is it causing more cerebral hypoperfusion in these patients? I think it's a really, really good question and very good remarks. Blood pressure management in patients in sinus rhythm, the studies show that you have to be very low to decrease the risk to be protective. So 110. And some studies show even less, 100. And we don't know in AF patients if we should use the same cutoff. And I agree with you. It's like over or under oral anticoagulants. We just need to be. But we don't know yet. And we don't know the right control. We don't know which is the best control for blood protection. And the second question is about obstructive sleep apnea. When we ask the specialists for guidance about CPAP therapy, they recommend CPAP mainly when there is severe obstructive sleep apnea. And when you have somebody with atrial fibrillation, and one of the concerns is obviously about recurrence of atrial fibrillation. And second is about declining the intellectual function. So should we as cardiologists be aiming for treatment even with, you know, mild to moderate sleep apnea or leave it to the sleep specialists who aim for CPAP only in severe sleep apnea patients? Yeah, that's another very good question. Because when you look at the studies on sleep apnea and cognition, they are all small studies and the results are, you know, not completely comparable between studies. I don't have the answer now. I hope to have the answer next year because we did the brain AF study in AF patients, in low risk AF patients that was presented in AHA last autumn. And what we did in our patient, 1,300 patients, we assessed sleep apnea at the beginning of the study and at the end, a new diagnosis at the end of the study. We looked at the MOCA scoring patient who had treated sleep apnea and those who were not treated. But I cannot discuss the result before next year. Great. Thank you, Lena, and thank you for the questions. Okay. We have run out of time. Thank you. The next is in our joint session between IHRA and the Canadian HRS, the next Canadian, Bill McIntyre, talking about the management of post-operative atrial fibrillation. Bill, please. Start. Okay. Honoured chairpersons, colleagues, thank you so much for joining, waiting for my presentation to load. So it's a funny thing, post-operative atrial fibrillation, and I hope to take you through a little bit some of the ways that I've been thinking about it in the last little bit. So I'm going to use a four-point framework. First, we'll talk a little bit about triggers. I want to think about oral anticoagulation both for a short-term strategy and for the long-term, and of course, talk a little bit about our concept of rate and rhythm control. So the first thing is talking about the trigger in surgery. So this is how I encourage you to think about new-onset atrial fibrillation that shows up for the first time in the setting of a potentially reversible trigger. In green, I've highlighted what I think is really the common pathophysiology. There's a systemic physiologic insult. It could be inflammation, autonomic stress, volume overload, or electrolyte disturbance. It could be different based on the exact condition and the surgery. And then there's an interaction with the surgery that was just done, right? If you think about medical illnesses like sepsis, probably none. In cardiac surgery, a surgeon has put a knife into the heart, and they've insulted the pericardium. It gets very inflamed. There's a long spectrum in non-cardiac surgery. Sometimes we work very close to the heart and the thorax. Sometimes we're very far away when we're on the knee. Think about these things. And also, think about in cardiac surgery, oftentimes there's a lot of arrhythmic substrate. Patients who have cardiac surgery often have an enlarged left atrium. We'll talk about that a little more. So before you go into some of the fancier and the tougher decisions, just think about managing pain, inflammation, optimizing volume status, electrolytes, and covering with antibiotics. You'd be surprised sometimes the only rate and rhythm control that you need in the short term is taking care of that acute underlying illness. So what about this question about rate or rhythm control? So I'd ask you this question if you have the audience response system, which is true. Is rhythm control or rate control better or neither? So it does come up with the live content. There we go. OK. So I guess it does give you 15 seconds to answer. If you've got your phones out and you've logged out of Instagram for a second to punch that in. And let's see what we come up with. 3, 2, 1. Awesome. OK. Rhythm control, better. You know you're at an EP meeting when everybody picks rhythm control. I like that. So the truth is, actually, neither has been shown to be better. This is data in non-cardiac surgery a colleague of mine, Dr. Mike Wang, led. So it found, yeah, of course, rhythm control was better at getting you back to sinus rhythm. But these are all observational studies that are really low quality. There is only a small RCT and really limited data for other outcomes. I show the plot here for mortality over the long term. And you can see that summary diamond at the bottom is right across the middle. What about cardiac surgery, where there's a little bit more data? There's about eight randomized trials and 1,000 patients. And if you look across the outcomes here on the bottom, length of stay, mortality, atrial fibrillation recurrence, absolutely no evidence of difference in the existing studies. Does that mean it doesn't matter? Well, I would submit to you that a lot of the data that we have is not really reliable. This is the landmark study that we have. It was the Cardiovascular Surgeon Clinical Trials Network in 2016 in New England. And you look at the four bullets I've put on their consort diagram. First, they actually enrolled patients in the study pre-op, which meant these are not patients that their clinicians felt uncertain about for rate and rhythm control. And this was their initial treatment strategy. You can imagine somebody with just a little blip of AFib versus somebody who had new onset AFib. But by looking at their clinical characteristics, you've been surprised that this is the first time it shows up in their life. But more importantly, on the third row, the crossovers, 26% and 24% in either arm. Friends, you have no chance of showing a difference in a clinical trial if you have that rate of crossovers. We want to aim for below 5%, maybe 10% at the most. But I think the challenge here is that this trial shows us that a trial like this is difficult to do. And I think that it's reasonable to go back to the drawing board to try to sort that out. In the meantime, you get to use your clinical judgment. So the next question I'll put up is, which of these do you think is true? Do you think amiodarone is effective for rapid cardioversion? Are IV beta blockers rapid and durable? Is IV diltiazem free of side effects? Or is nothing that I've written there true? Ah, everybody likes amiodarone. OK, effective drug for rapid cardioversion. So I actually would say that amiodarone is not an effective drug for rapid cardioversion. This is a great drug for maintaining sinus rhythm. It's a great drug for facilitating cardioversion. But it's actually slow. And it works slower than all of the class one drugs if you go head to head. I often have to remind friends that IV amiodarone infusion is a vasodilator, right? You may be using it for rate and rhythm control in the short term, but often you're changing and needing to start a vasopressor because of that. And remember, if you only use a bolus of IV amiodarone, you actually have an effect that lasts mere minutes. IV metoprolol is a great drug to proof of concept, but really it only lasts 10 or 15 minutes. If you see an effect you like with IV metoprolol, consider getting the oral in. Don't forget with diltiazem that it's a negative inotrope. And of course, digoxin, rarely used in the outpatient setting these days, is actually a reasonably attractive option for some short term rhythm control over a couple of days in the inpatient setting. So when I see post-op AFib and people are going faster, I often ask that question in blue, what would this patient's heart rate be if they were in sinus, right? They'd probably be in sinus tachycardia and maybe that's a reflection more of the underlying trigger in physiology. So let's get into the meat of things and talk about this conundrum of oral anticoagulation. My patient has had postoperative atrial fibrillation. Are they going to have a stroke? So I think the first thing you need to do when you see a patient with postoperative atrial fibrillation is make a short term oral anticoagulation strategy. And I think when you're there, on the rounds, on the hospital ward, you needn't commit to a long term strategy, right? You have a chance to see this patient a few weeks or maybe months down the road. They'll declare themselves. You can learn more about them. And the actuarial risk of a stroke over 90 days, particularly when they've just had surgery the other day, really isn't that low and it's hard to frame the risks and benefits. Who do I send home with oral anticoagulation? I think it's a spectrum as these things add up. If they spent more time in AF in the hospital, if they've actually been in and out of AFib and had multiple discrete episodes, if their left atrium's bigger, if their CHADS VASc score is higher, or if we actually had to cardiovert them, those are the things that lean towards making me want to cardiovert, or sorry, give oral anticoagulation going home. Now let's talk about the long term oral anticoagulation strategy. So this is from the 2024 ESCAF guidelines. So they give a class 2A, B indication saying that long term oral anticoagulation should be considered for both cardiac and non-cardiac surgery. And I think that this statement is pretty strong based on the data. I like to check the references that support any kind of a statement like this. And if you go to this document to the end, you'll see what these are are retrospective cohort studies based on administrative data. And I'll show you one of them. I think this was really a landmark in the field. This was by Ghialdini and colleagues. Dr. Healy was a co-author back in 2014. So they showed for patients who had post-op AFib after non-cardiac surgery, as opposed to patients who did not have post-operative atrial fibrillation, in blue, that hazard ratio was 1.3. So they were more likely to have a stroke. And their overall risk of stroke at one year, my apologies, I got that on there wrong, is 1.5%. Right? So it's sitting in there in kind of that gray zone for what we think about for worry. And then over here on the other side for cardiac surgery, the hazard ratio was 1.3, apologies, it should have been 2.0 on the other side. And that risk of post-operative atrial fibrillation patient having a stroke at a year was around 1%, something that I think the guidelines would actually consider low. So it's important to think about the context not only of saying, am I higher risk than a patient who didn't have post-operative atrial fibrillation, but also as what is my absolute risk of stroke and does it justify oral anticoagulation? So I like to think about this like this in the absence of randomized control trial data. And by the way, there are RCTs ongoing of oral anticoagulation for both post-op AFib in cardiac and non-cardiac surgery. So you get patients moving from the left to right that have this physiologic stress from surgery and they have new onset AFib. Now, they go back to sinus rhythm. That could be through time or it could be from the hand of the clinician. And then we're left with this clinical dilemma. Is the atrial fibrillation that we've seen here secondary, benign, meaning these patients are actually going to have the same prognosis as everybody else who had surgery but didn't have AFib? Or have we actually seen the first presentation of paroxysmal atrial fibrillation, an entity we know that is associated with a risk of stroke, heart failure, and death? And I think the simple question to ask is, these events are probably going to be mediated by recurrence of atrial fibrillation. So I think it makes sense for most people to look for more AFib rather than to commit them. So another poll question. So for patients who have post-op AFib, if you follow them, what proportion are going to recur? Nobody, everybody, or any of these three that are in between? OK. And everybody's sort of clustering around the middle. The trick here is it actually depends on the population and how you monitor them, right? But nobody picked none or all, so that's great. So we actually worked to do some systematic follow-up of patients with non-cardiac surgery paired with people who had medical illness. We went into the hospital and we did a prospective cohort study of people who had surgery or had conditions like sepsis and had AFib and then went home in sinus rhythm. We enrolled them in our study and then went back to that hospital ward and found people who were similar but never had AFib. And then we followed them up over a year with ambulatory ECG monitors. So and here's what we found. First, I'm going to show this number here, 5%. That was the rate of AFib detection in the control population. This is a little bit akin to AFib screening. So I think that that's what you might expect if you do a couple of holters. But in people who had post-op AFib, their rate of atrial fibrillation was 33%, a number that's high and, of course, unique to the population. The relative risk there was 6%. And if you look at the subset of non-cardiac surgery, maybe those proportions were even a little bit more exaggerated. What are the risk factors that predicted things? In green, left atrial volume. The bigger your left atrium, the more likely you were to have atrial fibrillation recur and follow up. And in red, troponin, higher troponin, actually more protective, maybe speaks to a reversible insult. And then finally, what about cardiac surgery? Well, surprisingly, despite the ubiquity of this condition, the data are a little scattered. Systematic review highlighted in blue on the left shows people who had an implantable loop recorder after post-op AFib. And we find that three months out, about 18% of them have recurrent AFib. And then in a year, it's about 30%. In green on the right, we recently reviewed our local clinic data. We see everybody systematically who has post-op AFib with a 14-day holter. We found 10% recurrence. And this was, again, predicted by left atrial volume. So I'll give you these as take-homes. One is, when you see post-operative atrial fibrillation, correcting the drivers, I think, often takes care of the AFib in the short term. Don't forget about the efficacy and pharmacokinetics of rate and rhythm drugs. And do your best as an arrhythmia consultant, which most of you are, to try to educate your frontline clinicians on this. I think it makes the most sense in the absence of RCT data to follow patients for AF recurrence for a long-term oral anticoagulation strategy. We can expect about one in three patients after non-cardiac surgery and one in 10 after cardiac surgery to recur. And left atrial size seems to be that best widely available predictor of who's likely to have atrial fibrillation recur. Thanks. Thanks, Bill. Any questions from the floor? Okay. Yeah, yeah, so, and I would say, for sure, and I think that's where there's a short-term strategy and there's a long-term strategy, right? We have a guideline statement that says post-op AFib, consider anticoagulation. There's way too much nuance that I think goes into the initial decision, and I think that we should wait and go forward, right? You think of the spectrum of patients that have post-op AFib, you imagine an 85-year-old with a large atrium who breaks her hip and a 22-year-old man who's in a motor vehicle collision trauma, and they both go into AFib, right? You have to use clinical judgment. Thank you, very nice presentation. So what is your, could you comment on the two strategies? So first, one in three will get AF on the long-term for cardiac surgery, at least. So do you do not anticoagulate, do the monitoring and then anticoagulate if you see AF, or do you start anticoagulation and stop it if you don't find AF later on? Because this is the two options, these are the two options that people are talking about. Absolutely, and I think we need to remember too, particularly in the short-term, the decision to anticoagulate is more of a shared clinical care decision model, right? You have a surgeon who's done the surgery who probably has some appreciation of the ongoing bleeding risk, and they've got skin in the game, right, as well, right? So based on, I think, the factors that you see on the short-term, you can make a decision about oral anticoagulation, and then I think it's important to reassess that, right? And I see patients in a clinic where other people have made the short-term decision, and I've been asked to reassess them to make the long-term decision. So what I think, I just think there's not enough data at that first snapshot to tell the person that this is your anticoagulation strategy for the rest of your life, particularly when we know the value and ease of documenting atrial fibrillation recurrence and truly phenotyping them. Nice presentation. I enjoyed it very much. Thank you. Thank you. Can you comment on the choice of anticoagulant? Yeah, so in cardiac surgery, in the short-term, if you talk to thrombosis colleagues and surgeons, there's a lot of places where warfarin remains the standard of care for oral anticoagulation in the first 90 days after cardiac surgery. Reason for this is the biochemistry is thought to be a little bit different, and I think we're all familiar with the Realign trial that used dabigatran, and they saw, among things, an increased incidence of pericardial infusions. There's lots of centers also that say, hey, whatever, it doesn't matter. I'm part of a large randomized trial now that's actually comparing the two, right? So I think in the first 90 days after cardiac surgery, the jury's absolutely out. I think that in non-cardiac surgery and then cardiac surgery beyond 90 days, it's just whatever is your normal oral anticoagulation approach. One more. One last one. Yeah, one more quick question. Sorry, I just was wondering, in my practice, I use a lot of insertable cardiac monitors to monitor these patients, and I was wondering, you presented data for both a patch and an insertable cardiac monitor. Yeah. Wondering what your thoughts were between the two and what you're using. Yeah, for sure. So in the jurisdiction where I work, insertable cardiac monitors aren't as readily available. If you have that as a strategy, you're absolutely going to get the totality of that patient's AF burden without missing anything going forward. And I think that we can lean back on what we know from subclinical atrial fibrillation, which my upcoming colleagues will talk about, to match them into that pathway. And as we learn more about atrial fibrillation burden and maybe where the threshold is if it's lower, that's the main tradeoff that you're getting between those two strategies. Great. Thanks, Bill. And good segue, because now it's time to battle here a little bit. We've got a debate coming up around subclinical atrial fibrillation. Should we anticoagulate or should we not? On the pro side, you're about to see Emma Svenberg from the Karolinska come up and give the pro side, to be followed by my co-chair, Helmut Purfeilner from Linz, to give the other side of the debate. So Emma, the floor is yours. Well, dear chairpersons, dear colleagues, it is indeed an honor to be here to start the debate on this topic. So let's see if I can also get my slides. Yes. Here we go. So I had a lot of thoughts going into this debate. So I'm going to debate the pro side against Helmut Purfeilner. For those who do not know Helmut, well, you know, he started in the European Heart Rhythm Association, became the president. He started running as a young person, became an Olympic runner. He started singing, joined a professional choir. To debate against Helmut is not a wise decision. It's more like a David towards Goliath situation, but I'm going to try to do my best. And why do I think this topic is important then? Well, stroke is still the second most common cause of death globally. Here in the US, there's one stroke every 40 seconds. And stroke is really a feared, feared disease if you canvass people globally. So it just comes after the loss of a family member, cancer, and then the third most feared event is a stroke. It's way up, which would very much confuse my children, the loss of your mobile phone. If we look at stroke epidemiology, and here's data from Sweden, we have quite good registry data, and we could see that up until the year 2010, we had a lot of ischemic strokes. And many of them were caused by AFib. And the proportion of AFib-related strokes, they were really going up. After 2010, we got the DOACs. And you could see that treatment of atrial fibrillation really brought a substantial change in the amount of AF-related ischemic strokes we had in Sweden. So for us, this made a huge difference in our health care systems. So we do know that treating clinical atrial fibrillation where patients have a lot of episodes of AF, they get cardioverted perhaps, we do know that they have quite a high stroke risk, right? But nowadays, we can monitor and monitor our patients for long, and we can pick up these very few episodes of atrial fibrillation, like through screening, with our own consumer-based devices, or through device detected, or through devices. So the question really is, how do we treat those? Well, of course, it's kind of easy to say that we should treat the comorbidities, right? But how about anticoagulation? It's not so simple, is it? So let's look at the prevalence then. Well, it's really common, if you have a device, to have device-detected AF. About 30% will have AF on their device. And the risk of stroke in patients with a device is increased, but it's not as high as in clinical atrial fibrillation. So there was really room for randomized controlled trials here, seeing if oral anticoagulants could improve the situation for these patients. And there has been two major trials, as we know, the ARTESIA trial and the NOAA-AFNET trial. The trials are quite similar. In a way, they both choose elderly patients. The average age in both studies are 77 or 78. They had an increased CHA2DS2-VASc score, and they all had device-detected episodes of atrial fibrillation. A difference between the studies was that in the NOAA trial, they also allowed episodes longer than 24 hours. In both studies, patients with device-detected atrial fibrillation were then randomized to a DOAC or a comparator drug. And the comparator drug in ARTESIA was aspirin, and in NOAA, it was aspirin if you were already on it, which a majority of the patients were. Otherwise, it was a placebo drug. A major difference, too, between these two studies are the number of patients included. Here, you could see that it's a much higher number of patients included in the ARTESIA trial, and also that the follow-up period was much longer in ARTESIA. And as a matter of fact, the NOAA-AFNET trial was terminated early due to futility and a signal towards harm. Also, when we look at the primary outcomes, they differed, both in the primary outcome variables, but also, of course, in the effect of the study. And we could see that the primary endpoint was positive for the ARTESIA trial in favor of oral anticoagulants, and not in the NOAA-AFNET trial. But there was also a signal towards bleeding, right? So if you're going to argue the pro side, I thought, this is quite worrisome, right? They have a lot of bleeding in the trial. So what sort of bleeding is this? And does this have a clinical consequence? So if you look into the bleeding outcomes, and this is data from the ARTESIA trial, you could see that actually fatal bleeding was very rare, actually rarer in numbers in the apixaban group compared to the aspirin group. And the clinical course of the patient was usually mined. So more than 80% just needed conservative or maybe supportive care. How about the strokes, then, the other side of the coin? What were they like? Well, if you look at the strokes that patients were protected from in the ARTESIA trial, you could see that there was a reduction in particular in the severe strokes, the strokes where you can no longer manage yourself independently, 49% reduction in those. So here, so you have to sort of think about the risks between these two. If you combine the trials, then, you could see that, well, they were quite similar. And when you combine them, you can see that the risk of ischemic stroke or all-cause stroke went down about 32% in both trials. But the risk of major bleeding was increased. However, lethal bleeding was not increased. So if you try to weigh the two, stroke or bleeding against each other, you could see that there were more bleeding events. But most of them had no long-term consequence for the patient. Whereas the stroke events, there you could see that there was really a reduction in the severe strokes here. So if you try to pinpoint down, then, who should really have oral anticoagulants? Who would benefit most in this population? I think we could look at some of the sub-studies and look at patients at high risk. So those with an increased CHA2DS2-VASc score of 4. If we look at that subgroup in the Artesia trial, we could see that this subgroup really benefited the most from oral anticoagulant therapy. And the number needed to treat in this group was 25. There was also no increased signal towards bleeding compared to the other groups. So the numbers needed to harm were much higher compared to the numbers needed to treat. Overall, there was a reduction of stroke risk of 56% in this group. How about other subgroups, then? Take patients with vascular disease. And this is a combined analysis from both Artesia and NOAAfnet, looking at patients with vascular disease defined as prior stroke, TIA, myocardial infarction, or peripheral vascular disease. Also here, we could see that the patients with clinical risk factor of vascular disease fared much better on an oral anticoagulant. And here, the risk reduction was 45% in that group. Finally, patients with a prior stroke, how did they do? Well, similarly to the other studies, they also fared much better if they had an oral anticoagulant. But another interesting group are those with longer episodes, right, of device-detected atrial fibrillation. We usually think about how long was the episode? Should I treat or should I not treat? But the studies here actually did not show a correlation with the stroke risk in these patients. So they did not seem to have an increased risk of stroke. However, as you could see in this substudy from the NOAA trial, you could see that patients with longer episodes of device-detected atrial fibrillation were much more prone to go on to have clinical atrial fibrillation. So the duration here more seems related to developing clinical atrial fibrillation than the risk of stroke. So how can we think about this? Well, I think about this as usual for my patients with clinical atrial fibrillation. I use a clinical risk score and I determine if they should have oral anticoagulants or not. And usually, most people use a breaking point of about 0.9% stroke risk to start treatment with a DOAC. Above that, that's usually sort of an intermediate risk of 1 to 2. And then above that, above 2%, we usually call it a high risk. So for most patients with clinical AF, they end up in the high-risk group. But CHA2DS2-VASc, if that's low, they can go down to a low risk as well. So for device-detected atrial fibrillation, they fall just in the group as a whole. They just are above the bar for an intermediate risk at about 1%. So that's a difficult group, right? But here, I also think, like we do with all other patients, I just assess the clinical risk factors. If they're high, if they have a high-chance VASc score, vascular disease, prior stroke, they end up in the high-risk group. And I think it's easy to decide on treatment then. So should we anticoagulate patient with device-detected AF? Yes, I think so. The rationale is stroke remains a major cause of death. You can reduce the risk of stroke with oral anticoagulants by about 30%. The severe strokes by almost half. And patients with a high-chance VASc score, vascular disease, prior stroke, they have a lot to gain from this treatment. We do know that patients have an increased bleeding risk. We have to acknowledge that. But it's mostly non-consequential. It won't carry with them for life. And then if you don't treat them now, they'll get AF anyway. So thank you. Thanks, Emma. I'm no longer sure which one of you is Goliath and which one's David. But Helmut, let's show us that world-class athleticism here with your con statement. The problem I have here, and thanks, Jeff, is that I like Emma a lot. Second, she's in the executive board of ERA, and I'm the president. So I could make a mistake in being too aggressive to her. And so my task is, on the one hand, easy, on the other hand, not easy. The easy thing is medicine and science and research. And I think that's the most important thing. The easy thing is medicine and science and research. And the other thing is to act politically. However, the easy thing is, remember what she has shown to you in one slide. She said in the years 2015, 2020, the last 10 years, the stroke risk went down. And she said all strokes went down. And then she brought another slide, or on the same slide, another figure where she had the DOACs on. I can tell you there's another study in Finland that the patients without, with AF and no anticoagulation had the same reduction in stroke. So it's not oral anticoagulation. It's better treatment. Remember that. To my slides, introduction is clear. One third has atrial fibrillation on device-detected tools. And if you have implanted an implantable device, of course, you will see everything over the observational period. You will not miss any AF. And what Jeff has done in 2012 is he showed in the ASER trial that there is an increased stroke risk in patients with SCUF, with subclinical atrial fibrillation, longer than six minutes, times 2.5. However, the magnitude, the risk of the stroke is higher. But does this relate to clinical consequences? That was the question. And do we reduce the stroke risk efficiently in the presence of oral anticoagulation, who also causes bleeding? And this was 2012. And in 2007, there was some additional analysis that if it's longer than 24 hours, there is higher risk you should do. If you then go to 2019, this was published in the LOOP study. So years later, you can see that in these 600 patients, with the natural history of subclinical atrial fibrillation detected by the implantable loop recorder, that in more than 680,000 monitoring days, the median burden in such patients is very, very low. AF longer than six minutes is in a third of patients. However, 22% will have no AF on their loop recorders if they have risk factors and a certain age. There is some progression to 24 hours or more. Most of them are asymptomatic, and the mean heart rate is around 20 beats higher. Does this mean that we should go for anticoagulation? And what they also could show is, if you have an episode more than 24 hours, there is only a sub-fraction of patients who will then have longer episodes. So it is not true that all of those having 24 hours or more will then have longer episodes. That's not true, according to what you see here. Only a sub-fraction will have that. Another one will not have it. A lot of them will not have that. So don't believe that. Okay, Atezia versus Noah Afnet. She showed everything. Thanks for that, Emma, so I can skip over that. However, look at the stroke rate. It's in the aspirin arm ischemic stroke 1%, 1.1%. So this is 1% stroke risk, and bleeding is 1.8. So it's almost doubling here, the bleeding rates under active drugs. Of course, you can speculate on what is the bleeding. It's more GI bleeding. It's less fatal bleeding. I'm okay with that. However, if you are a patient, you have a bleed, would you then continue? Would you? I'm not sure if I did. Emma was, of course, in her active mode in this discussion, but when she's more relaxed and she's doing some kind of scientific research, which I appreciate a lot, then she balances this risk and says, well, the duration of the ARRI itself should not always impose anticoagulation. So with the 24 hours, that is not true. She doesn't believe in that. And she believes much more in risk factor, competing risks, bleeding risks, and patient preferences. I'm okay with that, Emma. I think you should stay there. That's better. So if we go for the meta-analysis, stroke reduction, 32%, but 62% relative increase. If I go to you and I'm selling something and I show you something which has a benefit of 32% and a risk of 62, you will say, you're a crazy guy. I will never buy that. That is not really the thing I want to do. And the absolute risk is low, as you see here. So the bleeding risk clearly outweighs the stroke risk. And the stroke risk is low. And DOACs increase major bleeding risk, and bleeding can lead to serious outcomes. So it may outweigh the benefit. If you go to these two studies, I take two out of the three. The one is this Chad's VASC score study with Chad's VASC more than four. And this was from a tissue trial. So I don't want to risk too much if I criticize this trial. However, it's a subgroup, 24%, 25% of patient, of an essentially not very significant trial. The p-value for the interaction is non-significant. It's not confirmed in the other trial, the NOAA trial. And further analysis of atesia showed that the duration or frequency of the episode has no bearing on decision. Pah. Pah. Next trial. Yes, I'm with you. I'm with you. Emma, I think we can, politically, I think I'm now on the right track. I think we can convene on something, and we can consent on something. Because I think if you have vascular disease, then there is a higher rate of thromboembolic events, and there is still a risk of major bleed. But there seems to be a good relation. So to have patients with vascular diseases, stroke, myocardial infarction, peripheral vascular disease, and to anticoagulate them when they have device-detected AF makes sense for me. And for the others, we give aspirin. So in the question, aspirin versus DOAC, I think this makes sense. This makes sense. But not at the level of the question we're answering. And how can you predict something? Are we able to predict in a given patient that he or she has a stroke? Prediction is difficult, particularly when it involves the future. It's difficult. And now let's go to the strong arguments I have. The annual stroke risk in 2012, when ESSERT came out, was in SCAV, yes, with a CHA2DS2-VASc of 2 or 3, 3.8%. Do you know what the stroke risk nowadays is? 1%, one. So this is a massive decline in stroke risk, not only in those who have DOACs, but also in the others who do not have DOACs. Because we're treating patients better. We're treating their hypertension better. We have now better guidelines, stricter guidelines to treat hypertension. And of course, there are competing factors for stroke, not only atrial fibrillation. It is hypertension. And there is another risk in atrial fibrillation which we should think of, not only anticoagulation. The biggest risk is heart failure. Heart failure is the thing, and not stroke. So not to forget that. And the stroke risk has come down significantly. Significantly. And if you then look for the median device-detected AF and the annual stroke risk, you see the median device-detected AF in NOAA-AFNet was 2.8 hours, and in Atesia was 1.4 hours. In such short-duration episodes to give anticoagulation, well, I'm not sure. And looking at this paper, which I really like, from Becher and Kirchhoff, they show that nowadays, in 2025, stroke risk in red and percentage of AF burden in blue in device-detected AF, based on the loop study, we have 0.13% of AF burden. 0.13% of 100% burden we have in device-detected. That's what we are talking now. And we have a stroke risk of 0.5 to 1%. Not two, not three, not 4%, 0.5 to 1.1%. That's the stroke risk. So this might be my mom. She looks a bit different, but she's 86. My mom, if my mom had a device, and I would see that, she has no clinical AF, she has scuff. Six minutes, no. 24 hours, no. More than 24 hours, no, no. Why? So there is no impact on decision-making on all these factors. The only thing I would think about vascular disease, if she has something as a stroke in history, myocardial infraction, she doesn't have that, thank God, then I would give anticoagulation or aspirin in the case of no AF. So my conclusion is no routine anticoagulation for device-detected AF patients. There is a lower base rate stroke risk over the last decade without oral anticoagulation, not only with. Around 25% in one finished study. More in women, we treat women better. More in the elderly, without anticoagulation. The bleeding risk nowadays is still unchanged. There is no reduced risk of bleeding over the whole years now. So I think it doesn't warrant it. We need future research needed to identify high-benefit subgroups. I think we will need other tools for prediction, more based on other biochemical parameters that help us in predicting an event. And the clinical decisions must be personalized. I think vascular disease might help us out. Thank you. All right, so we don't have time for rebuttals, but we do have time for a few questions if Helmut and Emma would like to take them from their seats. And then at the end, we're gonna have a little vote, the Swedes versus the Austrians. So from the floor. Yeah, this debate was extremely enlightening, so thank you for arranging that. My question is that here you're debating the use of anticoagulant, and you mentioned significantly reduced risk of stroke compared to 20, 30 years ago when the original trials, early 90s. So my question is now, based on that information from 20, 30 years ago, folks are now jumping on to put in the left atrial appendage occluder devices based on the device-detected AFib. So any comments regarding that, you know, when you're talking about stroke risk being low? Yeah, I think very well taken. I think we should do a study in patients in such cohorts without any anticoagulation and without any device to see what's going on. Larry? Yeah, again, fantastic debate, and nobody died, so that was very good. So. I really like her. My question is actually in the group where you decide, wherever you throw your line, not to anticoagulate, about whether you really need ASA or whether no ASA would be the better option because ASA does carry a bleeding risk. From my perspective, I would probably go with no ASA unless they had an indication for it, and I would monitor closely as, although Helmut said that most don't develop AF, I think very many do, yeah. Helmut, do you have a comment? I think one of the things as Emma and Helmut both elucidate is the groups where there may appear to be benefit, you know, vascular disease, prior stroke, and even some of the high-chance vascular patients, they're gonna largely be on ASA anyway regardless. The only other point I'd make to the first point is we have to be a little bit careful when we're looking at different study patient populations, pacemaker population versus a screening population like loop, and we interpret risk factors from event rates. The actual systolic blood pressure at the entry of Artesia was about two to three millimeters of mercury lower than an assert. So ask yourself the question, does a two to three millimeter reduction in blood pressure reduce 2 3rds of the stroke risk? Or might we be seeing something like patient selection going into randomized trials? We cherry-pick patients for trials. Are they representative of the same patients that we see in our clinical practice? So chew on that. One last question, then the big vote. I enjoyed the presentation very much. So my question relates to the lack of temporal association between device detection of AFib and strokes. And that's how the concept came about that the strokes may not be from the AFib, but the fact that the AFib is an indicator of systemic disease. And in this regard, I have found it very, very difficult to understand the concept of atrial cardiopathy. It's such a nebulous entity. Is it real? And then can we conclude that in these patients, the stroke is not cardiac? The fact that Arcadia doesn't show significant benefit from anticoagulation in these patients? Well, I do think that Artesia showed significant benefit, actually, with oral anticoagulation. There is a primary outcome that's positive in the trial, even though the benefit is low. I do agree that the temporality is difficult, right? We can't see a very good temporality in the other cohort studies that have been done. But I do think that atrial fibrillation is a disease. You have a lot of other risk factors as well. So not all strokes are cardiombolic in the group, but overall, you see an effect on ischemic stroke in the trial. So I couldn't hear you. So patients who don't have a correlation, temporal correlation, can we say that these strokes, if they occur, are not cardiac? It is maybe embolic from elsewhere, from the vascular system? So let's say you have a patient that comes in for cardioversion. You do a transesophageal, actually, you see a thrombus. You say, okay, we're not gonna do your cardioversion today. Go home, we'll rebook you in two months. And six weeks later, they come back with a stroke. Was it caused by that, because it's not temporally associated? Do you think that thrombus that's added, it's uncertain, right? Like how long does a clot sit in the left atrium before it leaves? It's not so easy to put a fence around that and say, okay, this is temporal, this is not temporal. Like how long can a clot sit in the left atrium before it causes a stroke? Unclear, but. Well, I would also allude to the studies that we have in ESOS patients with embolic stroke of undetermined source, because the trials that we did on oral anticoagulation were negative. So that means that, of course, in these patients with unknown sources, we cannot attribute everything to atrial fibrillation and oral anticoagulation. Otherwise, the studies would have been positive, but they are not. Lots to learn still. One last question. Thank you, very, very good talks. My question goes to the dosing of NOACs or anticoagulation. As you presented very well, the bleeding risk is quite high. And in elderly patients, I find it very difficult to dose NOACs correctly, as the studies that assess the dosing were undertaken in patients younger than 80 mostly. So those patients over the age of 80 years are, to my opinion, very difficult to treat. So what's your take on that? Would you systematically go to lower dosing? Because these people are the frail ones that these are those that bleed more frequently. Yeah, for me, I would follow the recommendation by the drug prescription still. So I would reduce in case a renal function was low, for example, weight was low, creatinine high in a patient's apixaban. So I would just follow that recommendation generally. All right, we're a tiny bit over time, but we have the blue corner from Sweden, hands up for winner of the debates, and the red corner from Austria. So all for the blue corner, put up their hand. So come on. All right. And the red corner, right down the middle. So it looks like we'll be back doing it again next year. Thank you.

Atrial Fibrillation

Cognitive Dysfunction

Oral Anticoagulation

AF Ablation

Postoperative AF

Stroke Risk

Bleeding Risk

Subclinical Atrial Fibrillation

Individualized Patient Care

Heart Rhythm Society 2025